Inferact: A New Company from the Creators of vLLM ($150M Seed) (inferact.ai)

Inferact, a new startup launched by the creators of vLLM, has announced a $150 million seed funding round aimed at revolutionizing AI inference. As the most widely used open-source LLM inference engine, vLLM's mission is to evolve into a comprehensive AI inference solution, addressing the growing challenges posed by increasingly complex models and diverse hardware. The team emphasizes that current AI deployment requires specialized infrastructure, creating barriers for widespread access to advanced models. By streamlining this process, Inferact aims to make deploying state-of-the-art AI as straightforward as setting up a serverless database. The significance of Inferact lies in its commitment to keeping AI infrastructure open-source while enhancing performance and support for new model architectures. With vLLM already supporting over 500 model architectures and compatible with more than 200 types of hardware accelerators, the startup plans to deepen this ecosystem and drive collaboration between model and hardware vendors. This approach not only lowers the entry barrier for startups and researchers but also ensures that innovations flow back into the community, preserving the open-source ethos that has been pivotal to the LLM landscape. Inferact's proactive hiring strategy will bolster its endeavor to lead in the pressing area of AI inference infrastructure.